396 J . Org. Chem., Vol. 44, No. 3, 1979
Edgar, P e t t i t , a n d K r u p a
Synthesis of ~-(5-Chlor0-2-pyridyl)glycinel~ Mark T. Edgar,lb George R. Pettit,* a n d T i m o t h y
S.K r u p a
Cancer Research Institute and Department of Chemistry, Arizona S t a t e University, Tempe, Arizona 85281 Received M a y 1,1978
A4practical synthesis of the new amino acid ~-(5-chloro-2-pyridyl)glycine (2) was developed (Scheme I). The react ion sequence was initiated by condensing 2-chloro-5-nitropyridine with diethyl acetamidomalonate to afford diethyl (5-nitro-2-pyridy1)acetamidomalonate ( 5 ) . Reduction of the 5-nitro group and application of the isoamyl nitrite-carbon tetrachloride reagent provided diethyl (5-chloro-2-pyridy1)acetamidomalonate(8). Selective base hydrolysi:, followed by enzymatic hydrolysis furnished amino acid 2. number of steps to amino acid 2, t h e anion of benzylideneglycine ethyl ester was allowed to react with 2-chloro-&nitro~ y r i d i n e However, .~ t h e product was imidazole 6. Conversion of nitropyridine 4 to chloropyridine 8 was accomplished by way of diethyl (5-amino-2-pyridy1)acetamidomalonate (7). Reduction of 5-nitropyridine 4 t o 5-aminopyridine 7 was achieved by catalytic hydrogenation using platinum on carbon. Synthesis of 5-chloropyridine 8 proved t o be more difficult. Application of the classic Sandmeyer H reactionlo resulted in extensive decomposition. Thus, several 1 nonacidic methods were evaluated. Reaction of aminopyridine 7 with cupric chloride nitrosyl (generated from cupric chloride a n d nitric oxide)'l gave a n unsatisfactory mixture. Although one product appeared to be t h e 5-chloropyridine 8, t h e yield was quite poor. Next, isoamyl nitrite was allowed to react with aminopyridine 7 in the presence of carbon tetrachloride.12 T h e result was a 42% yield of 5-chloropyridine 8. 3 5 Attempted generation of DL-amino acid 2 from 5-chloropyridine 8 by acid hydrolysis gave 2-(aminomethyl)-5-chloropyridine (10). Variation of reaction conditions failed t o eliminate the decarboxylation. Mild base hydrolysisl3 of 5chloropyridine 8 followed by careful acidification of the reaction mixture a t 5 "C provided DL-cu-(acetylamino)-5chloro-2-pyridineacetic acid (9). T h e aceturic acid 9 proved to be somewhat unstable and was characterized as the methyl ester ll.l*Acid hydrolysis of aceturic acid 9 led to (amino13 methy1)pyridine 10. T h e required L-amino acid 2 was readily prepared from aceturic acid 9 using hog renal acylase I.15 griseoluteus. Because of considerable interest in 593A as a With the synthesis of L-amino acid 2 successfully comcancer chemotherapeutic d r ~ g , we ~ , began ~ t o investigate pleted, attention was focused on t h e feasibility of converting possible synthetic approaches. T h e first objective was t o deL-amino acid 2 to piperazinedione 13. Because of t h e usual velop a practical synthesis of the hitherto unknown amino acid ready dimerization of amino acid esters t o piperazinediones, ~-(5-chloro-2-pyridyI)glycine (2, Scheme I).4 For this purpose initial experiments were directed a t preparing a n ester dewe evaluated two readily available 2-chloropyridines as porivative of L-amino acid 2. Esterification procedures a t tential starting material for condensation with diethyl acet e m p t e d included thionyl chloride-methanol,16 p-toluenetamidomalonate t o afford a 2-substituted pyridine system.5 sulfonic acid-methyl acetate,17 perchloric acid-methyl aceSuitable degradation of the diethyl acetamidomalonate tate18 (instantaneous decarboxylation occurred in this case), moiety would t h e n lead t o amino acid 2.6 boron trifluoride-methanol,lg diazomethane,20a n d trimethyl T h e most expeditious route t o amino acid 2 appeared to be orthoformate.21 Each set of reaction conditions proved unvia 2,5-dichloropyridine; however, t h e 2-chloro group could rewarding. not be displaced by diethyl acetamidomalonate carbanion. In T h e esterification difficulties, particularly with diazoanother a t t e m p t , a stronger nucleophile was generated from methane, were attributed t o side reactions involving t h e primethyl hippurate using lithium diisopropylamine (LDA).' mary amine. T h u s , a t t e m p t s were made to block t h e amino Upon reaction of the hippurate nucleophile with 2,5-dichloposition of L-amino acid 2 with a tert-butoxycarbonyl16 (Boc) ropyridine a number of products were obtained. T h e 4-subgroup a n d by converting t o a Schiff base derivative.22 With stituted pyridine 3 was isolated in highest yield. Apparently the Boc approach, decarboxylation occurred during isolation the C-4 proton of 2,5-dichloropyridine was removed by the (acidic) to yield Boc-amine 12. Although the product was not strong base a n d underwent Claisen condensation with the fully characterized, decarboxylation also seemed to occur ester group of methyl hippurate.8 during Schiff base formation.23 At this point, the more reactive 2-chloro-5-nitropyridine was A t t e m p t e d dimerization of L-amino acid 2 by heating selected for further study. T h e condensation of diethyl acewithout solvent (under vacuum) a t temperatures u p to 250 "C tamidomalonate carbanion with 2-chloro-5-nitropyridine cleanly produced t h e (aminomethy1)pyridine 10. Use of a proceeded smoothly in dimethylformamide, furnishing pyrsolvent, for example, ethylene glycol, resulted in extensive idine 4. When 1,2-dimethoxyethane was used as solvent, formation of 1-(ethoxycarbonyl) -3-methyl-6-nitroimidazo[l,5- decomposition. Analogously, employment of dehydrating agents24to promote dimerization gave complex mixtures. alpyridine ( 5 ) became significant. I n a n effort t o shorten t h e
Recently, we reported a structural determination of the antitumor antibiotic: known as 593A ( 1)2from Streptomyces
0022-326317911944-0396$01.00/0 0 1979 American Chemical Society
J . Org. Chem., Vol. 44, No. 3, 1979 397 Scheme I
,
0
NHCOCH, 7
NMCOCH, 4
ld
6
C1
8
12
\
10
NHCOCH, 11
a [(CH,),CH)] ,NLi, HMPA, T H F , EtO,CCH,N=CHC,H,. b CH,CONHCH(CO,CH,CH,),, DMF, NaH. C Pd/C, CH,OH. d(CH,),CHCH,CH,ONO, CCI,. eNaOH, H’ a t 5 “C. fCH,N,. g Hog renal acylase I. hHC1, A. iHCl and/or A. i f k (CH,),COCON,, NaOH, citric acid.
A unique reaction for dimerization of amino acids to piperazinediones reported by Rosenmund and Kaiser25 involves synthesis of a n N-carboxyanhydride derivative (NCA) of a n amino acid followed by treatment with aziridine to afford the piperazinedione. Approaches t o t h e NCA derivative of Lamino acid 2 using phosgene again resulted in extensive decomposition, and this route was abandoned. An alternative approach based on first reducing amino acid 2 t o t h e corresponding piperidine derivative also proved t o be impractical d u e t o consistent hydrogenolysis of t h e 5-chloro substituent. Precedence for t h e facile decarboxylation of L-amino acid 2 has been observed most notably with c~-(Z-pyridyl)glycine,~ a-(4-pyridyl)glycine,26 a n d DL-c~-(2-thiazolyl)glycine.~~ Because of the instability of L-amino acid 2, other routes to 593A (1) are currently being studied. However, amino acid 2 appears to be of further interest in respect to biological properties. The L-amino acid 2 has been found t o significantly inhibit (ED50 = 3.8 hg/mL) growth of t h e P388 (murine lymphocytic leukemia) in vitro cell line.2s Table I summarizes t h e carbon-13 resonances for t h e pyridine derivatives synthesized in this study, Assignments were
made using off-resonance proton-decoupled spectra, singlefrequency decoupling experiments, and a survey of substituent effects in aromatic systems by Levy a n d Nelson.29
Experimental Section Purchased reagents were obtained from J. T. Baker, Mallinckrodt, Inc., Aldrich Chemical, M C D Manufacturing Chemists, Ventron, Div. of Thiokol (n-butyllithium), Fairfield Chemical (N-nitroso-Nmethylurea), and Columbia Organic Chemicals (2,5-dichloropyridine). All solvents were redistilled prior to use, and ligroin refers to a fraction boiling at -60 “C. Solvent extracts of aqueous solutions were dried over anhydrous sodium sulfate. Silica gel F-254 (0.25 mm) was used for thin-layer chromatography (TLC), and E. Merck silica gel (70-230 mesh) was used for column chromatography. TLC plates were visualized with UV light, ninhydrin spray, or bromocresol green spray. The IH NMR spectra (obtained with a Varian Associates XL-100 and Bruker WH-90 instrument) were provided by Dr. J. Witschel, while one of us (M.T.E.) recorded the ‘H NMR spectra using a Varian Associates T-60. Tetramethylsilane (Me&) was used as the internal standard. All 13C NMR spectra were recorded by Dr. Witschel using a Bruker WH-90 instrument. Infrared spectra were recorded using a Perkin-Elmer 237B spectrophotometer and optical rotations employing an O.C. Rudolph Model 80 polarimeter. Low-resolution mass spectra (70 eV) were provided by h h . E. Kelly using an Atlas
Table I. 2,5-Disubstituted Pyridine 13C NMR Chemical Shifts”
4. 7. 8
compd no. R
registry no.
11
2
1’
2’
3’
4’
5’
6’
2
3
4
5
6
165.96 167.39 166.51 170.06 171.79
70.25 69.41 70.03 56.96 58.98
169.34 169.24 169.28 170.06
22.76 22.95 22.69 22.85
161.31 143.82’ 153.90 153.09 151.66
126.10 124.25‘ 126.04 124.38 126.36
131.63 122.27‘ 136.44 136.96 138.76
143.89 142.98’ 131.89 131.92 133.10
143.56 135.08 146.81 148.34 149.10
~
4 7 8 11 2d
NO2 NH2 C1
67938-67-4 67938-68-5 67983-69-6 67983-70-9 67983-71-0
13.88 63.30 13.95 62.52 13.88 62.78 52.86
Assigned by single-frequency a In parts per million relative t o (CH&Si in CDC13. Assignments for C-2 and C-5 may be reversed. proton decoupling. Original data were relative t o dioxane (D20) and were converted t o the (CH3)4Si scale using 67.40 ppm.
398
J.Org. Chem., Vol. 44, No. 3, 1979
CH-4B spectrometer. A IKofler melting point apparatus was employed to determine melting points (uncorrected). Elemental analyses were determined by Spang Microanalytical Laboratory, Ann Arbor, Mich. 2,5-Dichloro-4-(1-a1xo-2-benzamidoethyl)pyridine (3). A 250-mL three-neck flask was equipped with a mechanical stirrer, nitrogen source, and rubber injection system. Dry tetrahydrofuran (50 mL), dry diisopropylamine (5.6 mL, 4.0 g, 40 mmol), and dry N,N,,V',N'-tetramethylethylenediamine(4.6 g, 40 mmol) were added; the solution was cooled to -78 "C, and n-butyllithium (2.4 M in heptane, 17.0 ml,, 40 mmol) was introduced via the rubber septum. Before methyl hippurate: (3.9 g, 20 mmol) in tetrahydrofuran (20 mL) was injected, the brown solution was stirred for 20 min. After stirring the bright yellow mixture for 1h a t -78 "C, a tetrahydrofuran (20 mL) solution of 2,5-dichloropyridine (3.0 g, 20 mmol) was added (through the injection port). As the mixture was allowed to warm to 25 "C, a deep red color appeared and water (100 mL) was slowly added. The organic layer was extracted with water (2 X 50 mL) and 1.2 M hydrochloric acid (?5 mL). The combined aqueous extract was adjusted to pH 2-3 and extracted with ether (2 X 100 mL). Removal of solvent from the ether extract gave a brown oil. The oil (-2 g) was dry-loaded on silica gel (10 g) and chromatographed on a wet-packed ligroinacetone, (4:l) silica gel (200 g) column. After elution with 1L of the same solvent, a pale yellow oil was recovered in the next 150-mL fractinn. Triturating the oil with ether caused crystallization (0.25 g, 3.6041, mp 11:)-116 "C). Recrystallization (twice) from chloroform-ether (-5 " C ) afforded colorless needles of dichloropyridine 3: mp 120.5-121 5 "C; IR (KBr) 3280,1710, 1630, 1520, 1355,1310, 1112,878,680CIT-'; 'H NMR (CDC13) 6 4.76 (d, 2 H, J = 6 Hz, CHz), 7.22 (broad s, 1H, NH. partially obscured by aromatic H, exchanges with D20), 7.36-7.54 (m. 3 H, phenyl H), 7.55 (s, 1 H, C-3 H), 7.74-7.91 (m, 2 H, phenyl H), 8.48 (s. 1 H, C-6 H); mass spectrum mle (re1 intensity) 312 ( < l ) ,310 (